Machine Learning set to create quality improvements in manufacturing processes
Machine Learning set to create quality improvements in manufacturing processes
Quality failures in fabricated metal products created using Computer Numerical Control (CNC) machining are being reduced, thanks to new AI Machine Learning software from experts at STFC Scientific Computing.
A critical element of the advanced, high-value manufacturing sector, CNC directs sophisticated machines to laser-cut, shape and finish high-quality, precision pieces efficiently.
The CNC sector globally is expected to grow to USD 115 billion by 2027.
However, quality issues can account for 25%-30% of a company’s turnover when materials scrappage costs are combined with the costs of staff rework, delays in shipments, and other expenses.
CNC is typically used to create pieces for industries such as aerospace, automotive and nuclear, where the margin for error is minuscule. An estimated 75% of geometrical and dimensional discrepancies in finished products are due to temperature fluctuations.
STFC Scientific Computing’s AI for Science team collaborated with Emerging Data Technologies (EDT) to develop machine learning software that enhances the calibration of CNC machine temperatures. The EDT project encompassed 15 CNC machines from various manufacturers, each of which behaves slightly differently in terms of temperature fluctuations.
“Significant sums of money are involved in subpar machining. An SME manufacturer in the precision manufacturing sector with a £10million turnover and a scrap rate of just 3% faces materials wastage of £300,000 per annum. Improvements of 50% in materials wastage would save them £150,000 a year.”
George Jones, Founder of EDT
